Learning Strategies and Self-Regulation in Blended Learning. The Case of Data Science Specialization

This study investigates the relationships between digital traces of self-regulated learning (SRL), self-reported study motivation and academic achievements in the blended Data Science course. Using the log data of students’ interactions with virtual learning environments, metrics to measure the components of self-regulated learning were developed. Based on the temporal patterns of students’ activity, a Latent Profile Analysis detected four distinct groups of students, who apply different learning strategies. Significant associations between trace-based measures of self-regulation, self-reported SRL and academic achievements were found using the statistical framework. With the help of conditional variable importance, it was revealed which metrics perform as the most important predictors of academic performance in blended learning settings. Self-reported self-efficacy was found to be the most important predictor, followed by behavioural patterns of self-regulation.